An Evaluation of χT Estimation Techniques: Implications for Batchelor Fitting and ε

Abstract

Techniques for the accurate estimation of the dissipation rate of temperature variance ?T from temperature microstructure measurements are synthesized and evaluated. The techniques focus on the treatment of contamination in the integration of the vertical temperature gradient spectrum. The essential improvements come from estimating ?T from the wavenumber domain in which the signal exceeds the noise and from reconstructing unresolved wavenumber contributions using the best-fit Batchelor spectrum. The improvement in the estimates of ?T results in better Batchelor fits and more robust estimates of the dissipation rate of turbulent kinetic energy ε. High correlations of independent measurements of ?T and ε from two collocated FP07 thermistors demonstrate the value of the techniques and highlight the improvements at the low end of the resolvable ranges of ?T and ε.